DI-drive
Decision Intelligence Platform for Autonomous Driving simulation.
DI-drive is an open-sourece application platform under OpenDILab. DI-drive applies Deep Learning method to decision, planning and controlling task in Autonomous Driving simulation, with high ease of use and low difficulty to hands-on. DI-drive designes Casezoo to make simulation environment closer to real driving. DI-drive develops Deep Learning driving policy with Pytorch and DI-engine and supports Carla and MetaDrive simulator.
DI-Drive provides tutorials for training agents in MetaDrive and shares well-trained models. For more information, please refer to the following resources:
Documentation: https://opendilab.github.io/DI-drive/index.html